Orval Fairbairn wrote:

Actually, filament winding would be a poor choice for spars, as the
filaments should run primarily parallel to the spar and be concentrated
at the top and bottom. You do need some in the webs, to handle shear
loads, but an "I" section is the most efficient. A tubular spar for a
wing is also a poor choice, as it concentrates a lot of its tensile
strength at its center, where it doesn't get much loading.

A mast is a different story, as it is expected to take similar bending
loads in all directions; a spar does not.

The spar in my Jodel is a one-piece box spar and is the only
spar in the wing. It takes the lifting and landing loads, the drag
loads, and the torsion loads. The washout is built into it. It's about
7" deep and 12" wide at the center. It has four lumber members, one in
each corner of the box, and the top two are larger than the bottom two,
since wood is stronger in tension than it is in compression. Plywood
diaphragms maintain the box shape and the whole thing is closed with
plywood skins with the outer plies at 45° angles to take advantage of
the maximum directional strength of the wood. The lumber members are
heaviest at the center and lightest at the tips. The dihedral is in the
outer panels, and the lumber members are spliced at those points.
Shrike had suggested building wooden airplanes with semi-skilled labor;
I don't think I'd want such a spar as this to be built by someone who
wasn't either thoroughly trained and apprenticed or who would be
eventually flying it. It's a most elegant piece of engineering and is
the most complex and time-consuming part of the whole airplane. One
foolish mistake in jigging or gluing could just about trash the whole
thing.

Dan